Good Things Come ln Analytical Packages
THERE'S A SEISMIC SHIFT shaking up the database market, and data warehousing
is at the epicenter. According to the "Database Market Zoos" report
issued last year by Forrester Research (Cambridge, MA), the insatiable demand
for more and better corporate information is causing database managers at Fortune
1000 companies to rethink their buying strategies.
While relational databases and OLTP (online transaction processing) systems
won't disappear any time soon, 77 percent of the database managers Forrester
interviewed said their investments would shift dramatically by 1999, away from
OLTP and its updates, inserts, and deletes and toward tools for understanding
data, i.e., data warehousing.
Forrester predicts that the transaction database segment will settle into stable
6 percent growth. Data warehouses and their smaller scale cousins, datamarts,
will go on a 40 percent annual growth tear, hitting near parity
with the transaction database market by 2001.
The reason is simple: Data warehousing makes old
data relevant to whole new categories of users. Ask an accounts payable
system what the average turnaround time for payment of invoices was this year
versus last year, broken out by invoice categories no way. But answering
that type of query is a piece of cake for a data warehousing tool.
Data warehouses accumulate information that enables
proactive decision making. For example, a data warehouse can not only
track inventory but also provide tools to manage inventory more precisely. The
system may be configured to forecast credit problems, for instance, so that
a customer can act upon the problems early enough to resolve them more easily.
Talk about a compelling sale.
"From an integrator's perspective, data warehousing
is a wonderful business opportunity," agrees Joe Brown, product manager
of data warehousing at Microsoft.
Then there's the inherent risk of delivering and
supporting OLTP based applications. When an OLTP system goes down, customers
get worked up. But when a decision support system goes off line
for a while, it may be inconvenient, but it doesn't stop business the way losing
order entry or accounts receivable does.
Most integrators have long lists of established
clients with systems they've been happily using for years. Data warehousing
tools can provide those customers with a whole new way to use all the data they've
collected. It wasn't long ago that the se sorts of solutions were priced out
of the reach of most middle market companies. But no more.
It's quite affordable now to setup separate systems
to implement data warehousing. But while it's a great opportunity for integrators,
the technology still has a rather steep learning curve.
Warehouse Not Required
By definition, a data warehouse is information that
has been optimized so that complex, multifaceted data can be retrieved much
more quickly than is feasible with the simple, indexed rows of data from relational
databases. The first data warehouses relied on custom-written routines to organize
data and present it back to users.
When that approach proved too expensive and unwieldy
for most implementations, the concept of departmental sized data warehouses a.k.a.
datamarts arose. Datamarts target the activity of particular departments
within an organization, are relatively inexpensive, are easy to sell, and can
be implemented fairly quickly. More important, the people who use them are key
business decision makers often those who approve budgets.
Integrators and ISVs have rushed to market with
prepackaged datamarts designed to serve every conceivable industry. The newest
wrinkle packaged analytical data warehouse tools provides nearly
limitless opportunities for integrators, because such tools go almost anywhere.
Stick them on top of an existing data warehouse
or datamart, and you get faster processing. But because many products include
internal data cubes (mini data warehouses of their own), useful results are
also made possible by simply grafting these tools onto clunky old relational
databases and OLTP systems.
Packaged data warehouse analytical tools vary widely in price and capability.
The cheapest desktop systems cost in the hundreds of dollars; enterprise client/server
products cost hundreds of thousands.
The tools have a variety of names, with multidimensional
analysis tools, OLA P (online analytical processing), and ROLA P (relational
OLAP) being the most widely used. Whatever the name, they all let customers
perform complex queries on many different simultaneous vectors.
Microsoft will be entering the fray when it introduces
its own OLAP server sometime next year. Expectations vary about the technical
merits of the Microsoft offering, but everyone agrees that Microsoft's arrival
in the field will fan the flames of a datawarehouse frenzy.
According to Mike Quimby, VAR partnership manager
at MicroStrategy, a Vienna, VA-based OLAP tool vendor, data warehousing gives
integrators a competitive advantage. Competition among integrators selling OLTP
solutions is intense.
"Most integrators and ISVs who provide OLTP type
solutions compete with four or five other companies with similar offerings.
0ur tools provide them with an easy way to add a new layer of value and provide
competitive differentiation," says Quimby.
Some markets have caught on faster than others. Any integrator that approached
a large retailer such ac Wall Mart and tried to sell a plain, vanilla
OLTP system without a data warehousing element would be shown the door
pretty quickly, Quimby says.
Other industries are following suit. In 1997 some
integrators began to see the telecom, financial, and healthcare industries express
serious interest in OLTP tools. Radiant Systems, for one, a 630 person
integrator headquartered in Atlanta, was a data warehouse pioneer. The
company recently began using MicroStrategy's OLAP software to provide better
analysis for packaged applications it offers to the petroleum, convenience store,
entertainment, foodservice, and automotive industries.
Scott Kingsfield, solutions director at Radiant,
thinks the new OLAP software has helped his firm stay a cut above the rest.
"Analysis tools haven't changed our basic approach, but they do provide
a new component to our overall solution."
In fact, OLTP has become Radiant's principle lead
into new customers. The technology speaks directly to the issues of greatest
interest to CEOs, because the tools make companies more strategically effective,
Kingsfield says. And the new tools let him go back to existing customers, to
sell a whole new line of products that help get more value out of their existing
Another category of warehouse tool currently being
hyped is data miners. These tools provide automated routines that search through
data organized in a warehouse, looking for patterns. "They can point you
toward areas that you should be addressing," explains Paul Hill, vice president
of strategic partners and marketing programs at Cognos Software in Burlington,
But Hill and other experts warn that whereas analytical
tools are already appropriate for a broad customer base, data miners are not.
"Data mining is currently at the same stage that OLAP was five years ago,"
cautions Hill. "It's relatively undefined right now."
Demonstrate Solid ROI
How do you know if a client needs one of these newfangled data warehouse
tools? Some integrators determine who among their customers needs to take copies
of printed reports they receive and retype the report data into a spreadsheet.
When you find that kind of activity on a large scale, you've found a prime candidate.
Once you've located good prospects, then it's Sales
101: Present a solution that clearly demonstrates a solid ROI, advises Kurt
Waltenbaugh, director of enterprise decision support for Retek, an integrator
and ISV in Minneapolis. Retek booked about $3o million in sales of million dollar plus
OLTP systems for large retailers last year.
The solution, branded the Retek Data Warehouse,
is an OLAP application that leverages MicroStrategy's toolset. The best documented
example of ROI, according to Waltenbaugh, comes from a Retek customer in the
UK, the Storehouse clothing chain.
"They went live last September and used our
solution to adjust their decision making process during the holiday season,"
he reports. "The new system paid for itself in four months' time."
No doubt Storehouse is having a happy new year.
Joe Devlin is a principal of Armadillo Associates, a software integrator headquartered
in Half Moon Bay, CA near California's Silicon Valley. You may contact him at